Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results
We present a detailed description of probabilistic methodologies based on Bayesian updating and structural reliability analysis for optimizing survey and/or inspection frequencies in a way that provides cost-effective solutions for ongoing integrity management or indeed for life extension or pressu...
Збережено в:
| Опубліковано в: : | Проблемы прочности |
|---|---|
| Дата: | 2009 |
| Автори: | , , |
| Формат: | Стаття |
| Мова: | Англійська |
| Опубліковано: |
Інститут проблем міцності ім. Г.С. Писаренко НАН України
2009
|
| Теми: | |
| Онлайн доступ: | https://nasplib.isofts.kiev.ua/handle/123456789/48440 |
| Теги: |
Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Цитувати: | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results / A. Francis, M. McCallum, C. Jandu // Проблемы прочности. — 2009. — № 5. — С. 36-54. — Бібліогр.: 4 назв. — англ. |
Репозитарії
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859604499905118208 |
|---|---|
| author | Francis, A. McCallum, M. Jandu, C. |
| author_facet | Francis, A. McCallum, M. Jandu, C. |
| citation_txt | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results / A. Francis, M. McCallum, C. Jandu // Проблемы прочности. — 2009. — № 5. — С. 36-54. — Бібліогр.: 4 назв. — англ. |
| collection | DSpace DC |
| container_title | Проблемы прочности |
| description | We present a detailed description of probabilistic methodologies based on Bayesian updating and structural reliability analysis for optimizing survey and/or inspection frequencies in a way that provides cost-effective solutions for ongoing integrity management or indeed for life extension or pressure uprating. A full description of the method is provided and the application of the techniques is clearly illustrated through case studies.
Детально описано ймовірнісні методики, що базуються на обновленні даних за Байесом і розрахунку надійності конструкцій. Методики використову ються для оптимізації діагностичного огляду і (або) частоти проведення локальних інспекцій трубопроводів. Це гарантує економічний розв’язок проблем забезпечення працездатності трубопроводів, подовження їх довговічності чи оптимізації тиску в них. Суть запропонованої методики й особливості її використання описано на різних прикладах.
Подробно описаны вероятностные методики, основанные на обновлении данных по Байесу и расчете надежности конструкций. Методики используются для оптимизации диагностических осмотров и (или) частоты проведения локальних инспекций трубопроводов. Это гарантирует экономичное решение проблем обеспечения работоспособности трубопроводов, продления их долговечности или оптимизации давления в них. Суть предложенной методики и
особенности ее применения детально описаны на различных примерах.
|
| first_indexed | 2025-11-28T03:06:38Z |
| format | Article |
| fulltext |
UDC 539.4
Pipeline Life Extension and Integrity Management Based on Optimized
Use of Above Ground Survey Data and Inline Inspection Results
A. Francis, M. McCallum, and C. Jandu
Andrew Francis & Associates, UK
УДК 539.4
Продление долговечности трубопроводов и меры по обеспечению
их работоспособности, основанные на оптимизированном
использовании результатов наземной диагностики и локальной
инспекции
А. Фрэнсис, М. Маккаллум, Ч. Джанду
“Эндрю Фрэнсис и партнеры”, Великобритания
Подробно описаны вероятностные методики, основанные на обновлении данных по Байесу и
расчете надежности конструкций. Методики используются для оптимизации диагностичес
ких осмотров и (или) частоты проведения локальних инспекций трубопроводов. Это гаранти
рует экономичное решение проблем обеспечения работоспособности трубопроводов, про
дления их долговечности или оптимизации давления в них. Суть предложенной методики и
особенности ее применения детально описаны на различных примерах.
К л ю ч е в ы е с л о в а : трубопровод, вероятностные методики, надежность конст
рукции, долговечность, оптимизация.
N o m e n c l a t u r e
- corrosion defect depth
- critical corrosion defect depth
- depth o f corrosion defect requiring repair
p (a , t age | K , ) - conditional defect depth distribution
- defect depth distribution
- growth rate distribution
- initiation time distribution
- time
- age o f pipeline
- initiation time
- expected number of corrosion defects
- expected number of coating holidays
- corrosion growth rate
-age
P0 (a> t age )
P o (K )
P 0(ti )
t
t age
t i
E c
E h
K
c
ar
© A. FRANCIS, M. McCALLUM, C. JANDU, 2009
36 ISSN 0556-I7IX. Проблемы прочности, 2009, № 5
Pipeline Life Extension and Integrity Management ...
N a - number of indications by survey only A
N b - number of indications by survey only B
N ab - number of indications by surveys A and B
N c - number of corrosion defects
N h - number of coating holidays
P0 (a r > t age ) - probability o f exceeding repair criteria
Vc - variance of number o f corrosion defects
Vh - variance of number o f coating holidays
a - ratio o f corrosion defects to coating holidays
$ A - probability o f false indication of survey A
$ B - probability o f false indication of survey B
t A - probability o f detection o f survey A
t B - probability o f detection o f survey B
Introduction . Oil and gas pipelines have now been operating in many parts o f
the world for periods in excess o f 40, or in som e places even 60, years. Operation
has largely been very successful w ith relatively few fatal accidents when compared
w ith the operation o f other types o f hazardous equipment. However, due to the
increasing age o f pipelines, more rigorous inspection and maintenance regim es are
becom ing increasingly important to ensure that the existing safety record is
increased or even improved.
There is a world-wide recognition that, in addition to other damage
m echanism s, external corrosion can pose a serious threat to the integrity o f buried
onshore pipelines and the focus on methods to prevent corrosion failures is
increasing. The primary means o f preventing corrosion is the application o f a
protective coating prior to installation. However, it is w ell recognized that breaches
in the coating do occur for several reasons and that the number o f breaches can
increase with time. For this reason, secondary measures are included such as the
Cathodic Protection (CP) system s to mitigate the likelihood o f corrosion growth
w hen breaches occur. Even so, the effectiveness o f CP system s is known to be
variable and intermittent and external corrosion occur. In v iew o f this, above
ground surveys such as Direct Current Voltage Gradient (DCVG) and Close
Interval Potential (CIPS) are periodically undertaken to identify areas o f coating
loss and locations o f active corrosion, respectively. Additionally, in-line inspection
(ILI) tools are used to detect the locations o f actual m etal loss. The objective o f all
surveys and inspections is to locate the presence o f corrosion defects in a tim ely
manner in order that repairs can be undertaken before failure occurs. Accordingly,
excavations are performed at the m ost likely locations and i f corrosion damage is
present and exceeds som e pre-determined criteria, a repair is performed.
The system s and activities described above are basic components o f an
integrity management plan (IMP). However, the effectiveness o f the IMP depends
on the frequency at w hich the surveys and inspections are undertaken and the
reliability o f the equipment used. Different combinations o f surveys m ay be used
with different frequencies. For instance, ILI is likely to be used less frequently than
above ground surveys. Moreover, for a significant number o f pipelines, ILI is not
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 37
A. Francis, M. McCallum, and C. Jandu
possib le resulting in total reliance on above ground surveys. The integrity
m anagem ent process based on above ground surveys alone has becom e known as
External Corrosion Direct A ssessm ent (ECDA).
Irrespective o f whether ILI, ECDA, or a combination o f the two is used, it is
important to ensure that frequencies at w hich the methods are em ployed are
sufficient to ensure that risk o f failure is acceptably low. The determination o f the
survey and inspection frequencies thus needs to take account o f the reliability and
accuracy o f the tools and also the current condition o f the pipeline. Probabilistic
techniques are best suited to this purpose.
A full description o f the method is provided and the application o f the
techniques is clearly illustrated through case studies.
Integrity management o f high pressure oil and gas pipelines is recognized
world-wide as the primary means o f ensuring that the pipelines are operated safely.
In basic terms, integrity management involves
(i) a identification o f the potential causes o f damage to the pipe-wall;
(ii) a determination o f the severity o f each type o f damage that can be
tolerated without causing failure com bined with
(iii) a means o f preventing the occurrence o f damage;
and/or
(iv) a means o f ensuring that i f damage does occur it can be tolerated.
For onshore pipelines the potential causes o f damage are generally the same
w orld-wide and include external interference, external/internal corrosion, stress
corrosion cracking and construction processes (welding). However, due to
differences in operating environments and construction techniques the relative
importance o f these potential causes m ay differ across the globe.
M eans for determining the severity o f the damage that can be tolerated are
available for all types o f damage and m odels that m ay be used for this purpose
appear in various design codes and defect assessm ent standards. Such m odels are
som etim es referred to as limit state functions.
For damage that occurs instantaneously (and w ill either cause failure or it
w on ’t) these functions can be used to place bounds on the steady state operating
regime. For instance recognising that internal pressure w ill affect the tolerability o f
m ost types o f damage such functions can be used to determine the maximum
allowable operating pressure. M anaging the pressure is thus one means o f ensuring
that damage caused by external interference can be tolerated.
On the other hand, som e types o f damage (i.e. corrosion and construction
defects) w ill not necessarily cause failure immediately but w ill becom e progressively
more severe w ith time and cause failure in the future. Tolerance o f this damage can
either be ensured by preventing the growth and/or detecting the damage and
rem oving and repairing it before it can cause failure. For instance som e assurance
o f the initial sizes o f construction defects is obtained from adherence to established
construction processes and survival o f the hydrostatic pressure test. The growth o f
such defects is controlled by controlling the magnitude o f pressure fluctuations. On
the other hand various steps are taken to prevent the occurrence o f external
corrosion. These include the application o f a protective coating and the application
o f a cathodic protection system. However, it is generally found that external
corrosion does occur and internal inspections and above ground surveys are used
38 ISSN 0556-171X. npo6n.eMH npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
on a period basis to detect it. I f corrosion is found the tolerabiUty o f it can be
assessed based on the magnitude o f the signal from the inspection/survey tool and
action can be taken im m ediately or in the future based on an assessm ent its severity
according to the lim it state function.
From the above, a number o f points emerges.
Firstly, the perceived severity o f the damage that can be tolerated depends on
the accuracy o f lim it state function and on the precision o f the knowledge o f the
basic parameters such as w all thickness, material grade and operating pressure.
The true magnitude o f any damage that has been found depends on the
accuracy o f the survey/inspection tools.
The likelihood o f detecting the various types o f damage depends on the
reliability o f the inspection and survey techniques. In general larger defects can be
more reliably detected. There is thus an increased likelihood o f detecting corrosion
damage i f it has been growing quicker or the inspection the interval between
inspections is increased. However, larger defects are more likely to cause failure. It
naturally follow s that prevention o f corrosion (and other types o f damage) is the
m ost steadfast means o f integrity management. However, accepting this is rarely
possible, the use o f reliable tools at some optimum frequency is best substitute for
this.
Another important issue is that, in general, there are more small construction
defects earlier in life than there are larger defects. This means that in early life
there are few defects requiring attention. However, as time increases and defects
grow the number o f defects requiring attention increases. This means that integrity
management becom es more expensive as time increases. Similarly, as coating
deteriorates w ith time the number o f corrosion defects increases and management
o f corrosion becom es more expensive. This leads to the concept o f econom ic life
o f the system. D epending on the revenue from transportation there necessarily
com es a time when the cost o f integrity management dictates that operation o f the
system is no longer econom ically viable and the life o f the pipeline com es to an
end.
The above narrative gives a brief description o f integrity management and its
relationship w ith asset (econom ic) life.
W hile integrity management and asset life have been described in principle,
detail has thus far been neglected. However, based on the brief description given
thus far it is clear that options for undertaking integrity management can be wide
ranging. In the sim plest form prescriptive integrity management can be undertaken
following the guidelines o f design codes. This w ill allow the operator to demonstrate
compliance with legislation. However, due to the conservatism and shortcomings
associated with prescriptive design codes, it likely that such an approach is
expensive, does not capture all safety issues and perhaps w ill not allow operation
beyond som e nominal design life.
On the other hand, risk-based approaches that take account o f m any aspects o f
uncertainty m ay lead to reduced costs and a safer operating regime and m ay allow
considerable extension o f the econom ic life. The downside o f the use o f these
techniques is that additional expertise is required and rigorous analytical tools are
required to demonstrate the validity o f the approach to regulators. N onetheless the
pay-back can be considerable in terms o f both revenue and safety.
ISSN 0556-171X. Проблемы прочности, 2009, N 5 39
A. Francis, M. McCallum, and C. Jandu
The aim o f this paper is to present and demonstrate robust techniques for
com bining the data from above ground surveys and ILI data to optimise the
integrity management process and to extend the econom ic life o f the a pipeline. A
particular focus is given to ECDA.
E xternal C orrosion D irect A ssessm ent. ECDA has been proposed as a
viable alternative to hydrostatic pressure testing and in-line inspection (ILI) for the
purpose o f managing the integrity o f high pressure pipelines [1]. Accordingly, an
EC D A standard is now in existence [2].
The essence o f EC D A is to com bine the results o f two or more above ground
surveys in order to establish a level o f confidence in the condition o f the pipeline.
R ecognising that all above ground survey techniques are subject to uncertainty,
bell-hole excavations are performed to investigate the presence o f damage at the
locations o f positive indications o f the surveys and to allow any necessary repairs
to be made using appropriate techniques. B y repeating the process at sufficiently
frequent intervals, the time dependent deterioration process due to external corrosion
can be effectively managed.
A central issue associated w ith the use o f above ground surveys is the
uncertainty in their effectiveness. Notably, not all defects w ill be detected and at
the locations o f som e indications no defect w ill be found. The characteristics o f the
surveys associated with these two issues are often referred to as the probability o f
detection (PoD ) and the probability o f false indication (PfI).
It thus follow s that the data obtained from any above ground survey requires
interpretation using probabilistic techniques. To this end, building on an outline
approach given in [3], a detailed method has recently been developed [4] for
determining the number o f defects that are likely to be present based on the results
o f the particular survey and the values o f the PoD and PfI. The method includes the
effect o f updating the distribution as information from excavations is acquired and
accordingly determines when a sufficient level o f integrity has been attained.
It is noteworthy, however, that such uncertainty is not confined only to above
ground surveys. Inline inspection tools are also prone to such behavior.
In v iew o f this, it follow s that techniques used combining different above
ground survey results can also be used to com bine above ground survey results
with ILI data.
L ife E xtension U sing above G round Survey R esults. The approach adopted
here is based on the concepts o f ECDA.
EC D A is a four stage process is adopted. The four stages are referred to as:
1. Pre-assessment.
2. Indirect assessment.
3. Direct examination.
4. Post assessment.
P re -a sse ssm e n t. During the pre-assessm ent stage, preliminary investigations
are undertaken to determine whether the use o f above ground survey techniques is
viable. For instance, i f a significant portion o f the pipeline is buried under hard
surfaces then DCVG w ould not be possible.
For the present purpose it is assumed that the surveys have been established,
i.e., DCVG and CIPS and hence no further discussion o f this aspect o f the
pre-assessm ent is required here.
40 ISSN 0556-171X. npodxeMbi npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
N onetheless, another function o f the pre-assessm ent is to establish the current
condition o f the pipe based on previously gathered data. This is discussed in more
detail later
I n d ir e c t A sse ssm e n t. The above ground surveys are undertaken during the
indirect assessm ent stage. The purpose o f these is to allow an initial updating o f the
prior distributions.
For the present purpose the two survey techniques are considered to be CIPS
and DCVG.
D ir e c t E x a m in a tio n . A bove ground survey techniques are subject to
uncertainty. The purpose o f the direct exam inations is to reduce uncertainty by
providing direct information on the actual condition o f the pipeline. The information
obtained form direct exam inations is used to im prove the confidence in the
performance characteristics o f the above ground surveys and consequently to
im prove the confidence in the know ledge o f the condition o f the pipeline. This
confidence is further enhanced by performing any necessary repairs during this
phase.
P o s t A sse ssm e n t. Once it has been established that the current integrity o f the
pipeline has been brought to an acceptable position, the data that have been
obtained are used to determine the likely increase in the number o f coating defects
and the likely growth rates at locations o f active corrosion. This information is
used to determine the allowable time period before the next assessm ent w ill be
necessary. In particular, for the present purpose, the future safe operational life is
determined.
G eneral Principle. The purpose o f this section is to outline, through the use
o f basic formulations, how the four phases o f the EC D A process are used to
system atically achieve the objectives set out above.
The detail associated with the stated formulations is given later.
P re -a sse ssm e n t. Based on the above, the prior condition o f the pipeline is
determined by the distribution o f the number o f corrosion defects, p c0( N c ), and
the defect depth distribution, p 0 ( a ). The latter w ill depend on the age o f the
pipeline, t age, the time, t i , at w hich corrosion growth com m enced and the growth
rate, K . Since the two latter quantities w ill be subject to uncertainty, the defect
depth distribution can generally be expressed as
t age rc
P 0 ( a , t age) = I f P ( a , t age IK , t i )P o (K ) P o( t i )d K d ti , ( 1)
0 0
where p 0(K ) and p 0 ( t t ) are the initial distributions o f K and t t , respectively.
The conditional probability, p ( a , t age | K , t i ), depends on the growth process.
The above distribution allows the probability, P ( a r , t age), that any given
defect w ill exceed the repair criterion to be determined and this quantity is given
by
P0 ( a r , t age ) = f p 0 ( a , t age ) d a , (2)
ar
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 41
A. Francis, M. McCallum, and C. Jandu
where all defects deeper than w ill be repaired. This is an important consideration
for deciding when to terminate the EC D A investigation.
I n d ir e c t A sse ssm e n t. Denoting the ILI run as survey A and the DCVG survey
as survey B, the m ost general outcom e w ill be N ab locations at w hich both
surveys give a positive indication, N A locations at w hich only survey A g ives a
positive indication and N b locations at w hich only survey B g ives a positive
indication.
The above quantities allow the expected number o f coating defects E H 0 and
the expected number o f corrosion defects, E c 0, to be determined. In general these
quantities w ill depend on the probability o f detection and the probability o f false
indication o f each o f the surveys. It thus follow s that E h 0 and E c 0 w ill have the
functional forms,
E H 0 = E H 0 ( t A 0 , £ B 0, 0 A 0 , 0 B 0 , a 0 ) (3)
and
E C 0 = E C0 (£ A 0, £B 0, 0 A 0 , 0 B0, a 0 ), (4)
where t a 0 , t B 0 , 0 a 0 ,and 0 B 0 are the probability o f detection o f survey A ,the
probability o f detection o f survey B , the probability o f false indication o f survey A ,
and the probability o f false indication o f survey B , respectively. The subscript ‘0 ’
denotes conditions prior to any information from excavations being obtained. In
the above a 0 denotes the ratio o f corrosion defects to coating defects. In order to
evaluate the expressions given above, it is necessary to determine a value for a 0 .
This can be achieved by finding the solution to the equation
E C0(£ A0 , £ B0 , 0 A 0 , 0 B0 , a 0 ) = a 0E H 0 (£ A0 , £B0 , 0 A 0 , 0 B0 , a 0 ) . (5)
Explicit reference to the dependence o f E H 0 and E c 0 on N ab , N a , and N b
has been omitted as these quantities are not subject to change.
The variance in the number o f coating defects, Vh 0 , and the variance in the
number o f corrosion defects, Vc 0 , m ay also be obtained from the above ground
survey results and consequently w ill take the functional forms
Vh 0 = Vh 0(£ a 0 , £ b 0 , 0 a 0 , 0 b 0 , a 0 ) (6)
and
VC 0 = VC 0 ( £ A 0 , £ B 0 , 0 A 0 , 0 B 0 , a 0 ). (7)
It follow s that distributions o f the number o f coating and corrosion defects can
be obtained and expressed in the general forms
P h 0 ( N h ) = P h 0 ( N h , E H 0 , VH 0 ) = p H 0( N n , £A 0 , £B 0 , 0 A 0 , 0 B 0 , a 0 ) (8)
and
p C0( N h ) = p C0( N h , E C 0 , VC0 ) = p C0 ( N c , £A 0 , £B0 , 0 A0 , 0 B 0 , a 0 ) . (9)
42 ISSN 0556-171X. npo6n.eMH npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
N u m b er o f C orrosion D efec ts E x ceed in g R e p a ir C riterion . The number o f
corrosion defects exceeding the repair criterion is related to the number o f
corrosion defects through
N r c = N c P ( a r , tage), (10)
from w hich it follow s that the distribution, p rc 0 , o f the number o f defects, N rc ,
exceeding the repair criterion is given by
1 ( N c
p rC 0 ( N rC ) _ d ̂ Ä p C 0 □ , ' 7 , £ A 0 , £B0, 0 A0 , 0 B0, a 0
\ P 0( a r , tage )V
( i i )
P 0 ( a r , ^ge )
The above distribution has an expectation, E rC 0 , and variance, VrC 0 , given by
E rC0 = E C0(CA 0 , CB0 , 0 A 0, ^ B0 , a 0 )P (a r , t age ) ( H )
and
VrC 0 = VC 0 ( C A 0 , C B 0 , 0 A 0 , 0 B 0 , a 0 )P ( a r , t age ) 2 ' (13)
The above provide an initial v iew on the number o f excavations that w ill be
required.
D ire c t E x a m in a tio n . U pda tin g the D istr ibu tion s o f N H a n d N c .Excavations
w ill generally be performed at som e or all o f the locations at w hich one or more o f
the surveys gave a positive indication. The outcome o f the excavation w ill confirm
whether or not a coating defect or a coating defect and a corrosion defect is
present. This w ill allow the distributions o f N c , N h , and a to be updated.
The findings at each excavation w ill either be a corrosion defect, a coating
defect w ith no corrosion or no coating defect.
This allows the values o f a 0 , CA 0 , CB0 , 0 A0 ,an d 0 B0 to be updated. The
nature o f the updating depends on both the nature o f the surveys and the nature o f
the findings and these issues are described in full detail later. However, at the
present stage in this paper, it suffices to say that the results o f the M e excavations,
w ill result in the revised values a m e , Ca m e , Cbm e , 0 a m e , and 0 b m e ■ The
locations that are selected for excavation w ill have an effect on these values,
however, the order in w hich the excavations are performed w ill not.
Based on the new values o f a m e , Ca m e , Cb m e , 0 a m e , and 0 b m e , the
distributions o f N H and N C can be updated giving
p H 0 Me ( N H ) = p H 0 ( N H , E HMe , v h m e ) =
= p H 0 ( N H , Ca m e , Cb m e , 0 a m e , 0 bme , a m e ) (14)
and
p C 0Me ( N C ) = p C0( N C , E CMe , v cme ) =
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N9 5 43
A. Francis, M. McCallum, and C. Jandu
- PC o( N C , £ a m e , £ BME , Ф a m e , Ф BME , a ME )• (15)
N ote from the above that the functional form o f the distributions has not changed;
the change is only to the characterising parameters.
In addition to causing the characteristic parameters to change, i f M H coating
defects are found (with or without corrosion) and M c corrosion defects are
found, then a further m odification to the distributions is appropriate resulting in
P hMkM k ( N H , ?a m e , ?BME , 0 a m e , 0 BME , a M E ) =
and
p H o ( N H , £ a m e , £ BME , Ф a m e , Ф BME , a ME ) H ( N h - M h )
/ p H 0( N H , £a m e , £BME , Ф AMe , ФBME , a ME ) d N H
M
p CMEME ( N C , £AM , £BME , Ф AMe , Ф BME , a ME ) -
p C0( N C 5 £AMe 5 £BME 5 ФAMe 5 ФBME 5 a ME ) H ( N C - M C )
x
/ p C0( N C 5 £AMe 5 £BME 5 ФAMe 5 ФBME 5 a ME )d N c
M
(16)
(17)
where H ( • ) denotes the H eaviside step function.
D istr ib u tio n o f N u m bers o f R em ain in g D e fe c ts . The numbers o f remaining
coating defects; N HR 5 and corrosion defects; N c R 5 are given by
N HR - N H - M H (18)
and
N c r - N c - M c 5 (19)
from w hich it follow s that the distributions o f remaining defects are given by
p HMEMER ( N HR 5 £AMe 5 £BME 5 Ф AMe 5 Ф BME 5 a ME ) -
p H 0( N HR + M H 5 £AMe 5 £BME 5 Ф AMe 5 ФBME 5 a ME ) (20)
- » (20)
/ p H 0 ( N HR + M H 5 £ AMe 5 £ BME 5 Ф AMe 5 Ф BME 5 a ME )d N HR
0
and
p CME M e R ( N CR 5 £ A M e 5 £ BME 5 Ф AM e 5 Ф BM E 5 a M E ) -
44 ISSN 0556-171X. Проблемы прочности, 2009, N 5
Pipeline Life Extension and Integrity Management
p C0 ( N CR + M C , Ça m e , ÇBME , 0 ame , 0 BME , a ME ) (21)
» ■
f p C0 ( N CR + M C , Ça m e , ÇBME , 0 a m e , 0 BME , a ME )d N CR
U pda tin g the D istr ib u tio n o f a. Follow ing the M E excavations and the
discovery o f M c corrosion defects, it is also possible to update the corrosion
defect depth distribution. This is done by updating the distributions o f K and t i .
The joint distribution o f K and t i is updated using the expression it
Mr
n P ( a j 5 t age I t i 5 K ) P k 0 ( K ) P ti 0 ( t i )
j =1
p m e ( k ■ '• ) = -------------------------------------------------------------■ (22)
//n P ( a j , t age | t t , K ) P k o (K )P tio ( t i )d K d ti
0 0 j =1
where the a t values ( j = 1 to M c ) are the measured depths o f the corrosion
defects found at the excavations. I f follow s that the updated defect depth distribution
is given by
tage X
PM E ( a , t age ) = // P ( a , t age1 K , t i ) PM E (K , t i )d K d ti . (23)
0 0
Naturally, the above distribution w ill depend on the results o f excavations that
have been undertaken, however, it w ill be independent o f the order in w hich the
excavations are undertaken.
D istr ib u tio n o f N u m b er o f R em ain in g C orrosion D efec ts E x ceed in g the R eP air
C riterion . The probability that any single defect depth w ill exceed the repair
criterion after M E excavations have been undertaken is given by
P ME ( a r , t age ) =/ P M E ( a , t age )d a (24)
The number o f remaining corrosion defects exceeding the repair criterion, N rc R ,
is then given by
N rCR = N CR P ME ( a r , t age ) . (25)
It then follow s that the distribution, P,-cMeM eR ( N rc R ), o f the number o f remaining
defects, N rcR , exceeding the repair criterion is given by
1
p rCME Me R ( N rCR ) _ P (a t ) p CME Me R
P Mp ( a r 5 t age )E \~r? -age; y P Me ( a r 5 t age )
N rCR
(26)
0
ar
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 45
A. Francis, M. McCallum, and C. Jandu
The above distribution has an expectation, E tcMeM eR , and variance, VrcMEMER ,
given by
E rCMEMER = E CMe Me R ( a r , {age) (27)
and
VrCMEMER = VCMe MeR P Me ( a r , {age ) , (28)
respectively, where E c m m er and VcMeM eR are the expectation and variance
associated with P cMeM eR ( N c ).
The general principle outlined above applies to any E C D A m ethodology and
therefore represents a general framework for reference.
E xam ple. The intent o f this section is to show, using a simple exam ple, how
the m ethodology can be used as part o f an integrity management program to extend
the pipeline life.
P re -a sse ssm e n t. The pipeline parameters used in this example are given in
Table 1. It should be noted that since the m ethodology described in this paper
requires only know ledge o f the repair depth, there is no requirement to detail all
the pipeline parameters.
T a b l e 1
Pipeline Parameters
Parameter Value
Commissioning year 1970
Current year 2006
Diameter 219.1 mm
Wall thickness 6.35 mm
Grade X42
Pressure 5.5 MPa
Repair depth 30% wall thickness
Number of corrosion defects 10
In order to calculate the defect depth distribution, the distributions in Table 2
have been used to m odel the uncertainty in the above three parameters.
The distribution o f initiation depth is assumed to represent those defects that
are just becom ing detectable, i.e., smaller defects are not considered to be defects.
This assumption is acknowledged as being subjective; however, it suffices here for
the purpose o f illustration.
The initiation time is perhaps subject to m ost uncertainty in practice and
indeed this quantity has rarely been isolated and indeed m ay have an effect on
reported measured growth rates. It is acknowledged that more focus is needed on
the determination o f this quantity, and the values used below, w hilst not unrealistic
are used only for illustration purposes here.
46 ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
T a b l e 2
Distributions Used in Defect Depth Distribution
Factor Distribution Mean Standard deviation
Initiation depth Weibull 0.5 mm 0.5 mm
Initiation time, ti Normal 18 years 4 years
Growth rate, K Normal 0.2 mm per year 0.025 mm per year
U sing the data given above, the defect depth o f the 36 year old pipeline was
determined using Eq. (1) and the results are shown graphically in Fig. 1. It is seen
from Fig. 1 that the m ean depth increases from 0.5 mm to about 1.75 mm in the 36
year period based on an average growth rate o f 0.2 mm per year. A t first glance this
result appears anomalous. However, it is important to note that not all defects w ill
have been growing for the full 18 year period. Those defects introduced towards
the end o f the 36 year period w ill have been growing for a very short time and this
is reflected in the results given in Fig. 1.
Fig. 1: Defect depth distribution (solid line correspond initiation distribution; dashed line - 2006
distribution).
U sing the defect depth distribution and Eq. (2), it is possible to make an initial
assessm ent o f the condition o f the pipeline. The probability that any given defect
w ill exceed the failure criterion, given in Table 1, is 0.235. This means, using (12),
that the expected number o f corrosion defects exceeding the repair criterion is
around 2.35.
Structural Reliability A nalysis (SRA) can be used to determine the future
failure probability for the pipeline. Details o f the approach used can be found in
[3]. The failure probability for the pipeline is shown in Fig. 2. It can be seen that in
2010 (40 year design life) the failure probability is marginally greater than_3
1.0-10 per km. Based on AFAA’s experience o f conducting similar studies this
figure is considered high, although a more extensive study w ould look at the risks
associated with this failure probability.
I n d ir e c t A sse ssm e n t. The number o f indications from the Corrosion Survey
and DCVG survey is given in Table 3.
From Table 3, it can be seen that there were five areas where both the
Corrosion Survey (survey A ) and the DCVG (survey B) were positive, no areas
where only the Corrosion survey was positive and tw enty-five areas where only the
coating survey was positive.
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 47
A. Francis, M. McCallum, and C. Jandu
T a b l e 3
Above Ground Survey Results
Indication type Number of indications
N AB 5
N a 0
N b 25
1.00E+00
1.00E-02
1.00E-04
1 00E-06 -
1.00E-08 -
1.00E-10 -I----------- ,------------,----------- ,----------- ,----------- ,------------,----------- ,-----------
2004 2006 2008 2010 2012 2014 2016 2018 2020
Fig. 2. Future failure probability based on pre-assessment information.
For the purposes o f this example, the m ean values o f the probability o f
detection (£) and probability o f false indication (ф) o f the above ground surveys A
and B are given by Table 4.
T a b l e 4
Above Ground Survey Characteristics
Probability ILI (survey A) DCVG (survey B)
Probability of detection 0.76 0.84
Probability of false indication 0.10 0.11
These values are based on previous comparisons between indicated and
observed results on earlier projects. However, it is likely that different values w ill
be appropriate to other situations and hence the values given here should be
regarded as indicative only. Notwithstanding this, the updating w ill cause a
migration o f the initial starting values towards the true values in any given
situation.
The distributions for the number o f corrosion defects and the number o f
coating holidays are shown in Fig. 3.
The expected number o f corrosion defects after the indirect assessm ent is 6.6.
U sing (12), the probability o f a single defect exceeding the repair criterion is 0.235
and the expected number o f corrosion defects exceeding the repair criterion is 1.55.
So after the indirect assessm ent there has been a decrease in the number o f
corrosion defects exceeding the repair criterion, this is due to the updating o f the
number o f corrosion defects based on the above ground survey results, taking
account o f the accuracy o f the above ground techniques.
48 ISSN 0556-171X. Проблемы прочности, 2009, № 5
Pipeline Life Extension and Integrity Management
4.00E-01
|> 3 00E-01
™ 2 OOE-O I n
2 1.00E-01
OOOE+OO
Fig. 3. Distribution of corrosion defects and coating holidays after indirect assessment (dots
correspond coating holidays; solid line - corrosion defects).
D ir e c t E x a m in a tio n . Excavations, based on the results o f the above ground
surveys, are undertaken during the direct examination stage. The purpose o f these
is to allow us to update our v iew on the accuracy o f the above ground surveys, the
number o f corrosion defects on the pipeline, the size o f corrosion defects on the
pipeline and hence the number o f corrosion defects that exceed the repair criterion.
For this example, it has been assumed that twenty excavations have been
undertaken. Five o f the excavations were carried out at areas where both surveys
were positive; at each o f these excavations, a coating holiday and corrosion were
found.
Fifteen excavations were carried out at sites where the coating survey was
positive; at each o f these excavations a coating holiday with no corrosion was
found. The sequential changes to the characteristics o f the above ground surveys
are shown in Table 5.
From Table 5, it can be seen that the probability o f detection o f the corrosion
survey (£ A ) increases (i.e., survey is more accurate) from 0.76 to 0.79, this is due
to the fifteen excavations where the corrosion survey was negative and no
corrosion was found. The probability o f false indication o f the corrosion survey
( $ a ) and the probability o f detection o f the coating survey (£B ) are unchanged as
there have been no excavations where the corrosion survey w as positive and the
coating survey w as negative. In the final column o f Table 5, it can be seen that the
probability o f false indication o f the coating survey (<p B ) decreases (i.e., survey is
more accurate) from 0.11 to 0.09. Again, this is due to excavating fifteen locations
where the coating survey w as positive and a coating holiday w as found at each
excavation.
The change in number o f corrosion defects and the number o f remaining
corrosion defects is shown in Table 6 and Fig. 4.
From Table 6 and Fig. 4 it is seen that the overall expected number o f
corrosion defects decreases slightly. This is due to an increase in the probability o f
detection o f the Corrosion Survey, i.e., the corrosion survey is less prone to
m issing defects.
In this case, the expected number o f remaining unknown corrosion defects
again decreases in direct accordance with the total number; the difference being
generally due to the 5 known corrosion defects found at sites where both surveys
were positive.
Since corrosion defects were found during the excavations, it is also possible
to update the initiation time and growth rate distributions, w hich in turn allows us
to update the defect depth distribution. The depths o f corrosion defects found
during the excavations are shown in Table 7.
Number
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 49
A. Francis, M. McCallum, and C. Jandu
T a b l e 5
Effect of Excavations on Survey Characteristics
Excavation £b $ B
0 0.76 0.10 0.84 0.11
1 0.76 0.10 0.84 0.11
2 0.76 0.10 0.84 0.11
3 0.76 0.10 0.84 0.11
4 0.76 0.10 0.84 0.11
5 0.76 0.10 0.84 0.11
6 0.76 0.10 0.84 0.11
7 0.76 0.10 0.84 0.11
8 0.77 0.10 0.84 0.10
9 0.77 0.10 0.84 0.10
10 0.77 0.10 0.84 0.10
11 0.77 0.10 0.84 0.10
12 0.78 0.10 0.84 0.10
13 0.78 0.10 0.84 0.10
14 0.78 0.10 0.84 0.10
15 0.78 0.10 0.84 0.10
16 0.78 0.10 0.84 0.10
17 0.79 0.10 0.84 0.09
18 0.79 0.10 0.84 0.09
19 0.79 0.10 0.84 0.09
20 0.79 0.10 0.84 0.09
Change in Expected Number of Corrosion Defects
0 5 10 15 20 25
E x c a v a t i o n
Fig. 4. Effect of excavations on the expected number of corrosion defects and number of corrosion
defects remaining (dots correspond expected number of corrosion defects; solid line - expected
number of remaining corrosion defects).
The effect o f the corrosion defects found during excavations on the defect
depth is shown in Fig. 5. From Fig. 5 it can be seen that the there is a gradual shift
to the left, i.e., excavations are saying that the depth o f corrosion defects are not as
deep as was thought during the pre-assessment. This could be due to the initial
corrosion growth rate being too high, the time corrosion defects initiate being too
early or a combination o f both o f these.
50 ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
T a b l e 6
Effect of Excavations on Expected Number of Corrosion Defects
and Number of Corrosion Defects Remaining
Excavation EC ecr
1 6.59 5.39
2 6.59 4.49
3 6.59 3.29
4 6.59 2.42
5 6.59 1.74
6 6.57 1.72
7 6.55 1.70
8 6.53 1.69
9 6.50 1.67
10 6.48 1.65
11 6.47 1.64
12 6.45 1.62
13 6.43 1.60
14 6.41 1.59
15 6.39 1.57
16 6.38 1.56
17 6.36 1.55
18 6.34 1.53
19 6.33 1.52
20 6.31 1.51
T a b l e 7
Depth of Corrosion Defects Found During Excavations
Excavation Corrosion depth (mm)
1 1.2
2 1.0
3 0.9
4 1.1
5 1.0
The effect o f the above ground surveys and the excavations on the expected
number o f corrosion defects exceeding the repair criterion is shown in Fig. 6.
From Fig. 6, it can be seen that there is a steady drop in the expected number
o f corrosion defects greater than the repair criterion.
Based on the expected number o f corrosion defects greater than the repair
criterion, a decision can be made on whether further excavations are required. For
the purposes o f this example, it is assumed that excavations are necessary until this
value is 5% or less (highlighted by the bold line in Fig. 6). Thus on this occasion, it
w ould be appropriate to stop excavating after 5 excavations.
ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5 51
A. Francis, M. McCallum, and C. Jandu
1.00
>, 0.80 4—'
I 0.60
to
■g 0.40
uL o.20
0
f / . 'A
jti >
r
/
0 1.00 2.00 3.00 4.00 5.00
Depth (mm)
........... Excavation 1 ------------Excavation 2
--------- : Excavation 3 ----------- - Excavation 4
-----------Excavations ------------Prior Distribution
Fig. 5. Effect of excavated corrosion defects on the defect depth distribution.
Fig. 6. Effect of excavations on the expected number of defects greater than the repair depth.
Fig. 7. Comparison of pre- (solid line) and post-assessment (dashed line) failure probability.
P o s t A s s e s s m e n t a n d L ife E x te n s io n . The post assessm ent stage uses the
updated distributions for the defect depth and number determined in the indirect
and direct examination stages to determine an updated failure probability for the
pipeline. This failure probability can be used to determine the date for the next
integrity assessm ent or in this example to be used to determine the length o f life
extension possible.
SRA was used to determine the failure probability based on the updated
distributions.
The updated failure probabilities are shown in Fig. 7 for the time period
between 2006 and 2020. It is im m ediately obvious that using the information
gathered from the above ground surveys and the excavations that the failure
52 ISSN 0556-171X. npoôëeMbi npounocmu, 2009, N 5
Pipeline Life Extension and Integrity Management
probability due to external corrosion has reduced to negligible values up to around
2010. Furthermore, assuming an acceptable failure probability o f around 1.0 • 10_4
per km, it can be seen that the failure probability remains acceptably low until
around 2020.
Based on the above, the EC D A m ethodology has shown that probability o f
failure remains acceptable till around 2020. The implications o f this outcome are
that it is possible to schedule future integrity assessm ents, but it is also viable to
consider extending the life o f the pipeline beyond the 40 year nominal design life.
C o n c l u s i o n s
1. The follow ing methods have been presented:
- a method for determining the corrosion defect depth distribution based on
variable growth rate, variable growth time, and time dependent introduction o f
corrosion defects;
- a method for determining the expected number o f coating defects, the
expected number o f corrosion defects, the associated variances, and hence the
distributions, based on the results from a coating survey and a Corrosion survey;
- a method for updating the corrosion defect distribution based on the
measurements made at excavations;
- a m ethod for sim ultaneously updating the probability o f detection and the
probability o f false indication o f each survey technique based on the results from
excavations;
- a method for updating the distributions o f the numbers o f coating and
corrosion defects based on the results from excavation;
- a method for determining the distribution o f the number o f remaining
defects follow ing excavation and repair;
- a method for determining the expected number o f defects that w ill exceed
the repair criteria and the confidence limits on this quantity.
2. A sim ple example showing how the method is used to extend the life o f
ageing pipelines has been presented.
Р е з ю м е
Детально описано ймовірнісні методики, що базуються на обновленні даних
за Байесом і розрахунку надійності конструкцій. М етодики використову
ються для оптимізації діагностичного огляду і (або) частоти проведення
локальних інспекцій трубопроводів. Це гарантує економічний р озв’язок
проблем забезпечення працездатності трубопроводів, подовження їх довго
вічності чи оптимізації тиску в них. Суть запропонованої методики й особли
вості її використання описано на різних прикладах.
1. US C ode o f F e d era l R egu la tion s (CFR) T itle 49. “T ransportation o f H azardou s
L iq u id s b y P ip e lin e ,” Part 195. W ashington, DC.
2. N A C E R eco m m en d ed P ra c tic e R P 0 5 0 2 -2 0 0 2 , P ip e lin e E x tern a l C orrosion
D ir e c t A ssessm en t.
ISSN 0556-171X. Проблеми прочности, 2009, № 5 53
A. Francis, M. McCallum, and C. Jandu
3. A. Francis, M. Gardiner, A. G oodfellow , et al., “A systematic risk and
reliability-based approach to integrity management o f piggable and non-
piggable pipelines,” P ip e lin e In tegrity a n d S afety Conference, Houston (2001).
4. A . Francis and C. Jandu, “O ptim ized E C D A based on sim ultaneous
interpretation o f multiple above ground survey techniques and identification
o f required excavation sites using a probabilistic m ethodology,” N A C E 2006,
San D iego (2006).
Received 05. 01. 2009
54 ISSN 0556-171X. npo6n.eMH npounocmu, 2009, N 5
|
| id | nasplib_isofts_kiev_ua-123456789-48440 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 0556-171X |
| language | English |
| last_indexed | 2025-11-28T03:06:38Z |
| publishDate | 2009 |
| publisher | Інститут проблем міцності ім. Г.С. Писаренко НАН України |
| record_format | dspace |
| spelling | Francis, A. McCallum, M. Jandu, C. 2013-08-19T15:37:40Z 2013-08-19T15:37:40Z 2009 Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results / A. Francis, M. McCallum, C. Jandu // Проблемы прочности. — 2009. — № 5. — С. 36-54. — Бібліогр.: 4 назв. — англ. 0556-171X https://nasplib.isofts.kiev.ua/handle/123456789/48440 539.4 We present a detailed description of probabilistic methodologies based on Bayesian updating and structural reliability analysis for optimizing survey and/or inspection frequencies in a way that provides cost-effective solutions for ongoing integrity management or indeed for life extension or pressure uprating. A full description of the method is provided and the application of the techniques is clearly illustrated through case studies. Детально описано ймовірнісні методики, що базуються на обновленні даних за Байесом і розрахунку надійності конструкцій. Методики використову ються для оптимізації діагностичного огляду і (або) частоти проведення локальних інспекцій трубопроводів. Це гарантує економічний розв’язок проблем забезпечення працездатності трубопроводів, подовження їх довговічності чи оптимізації тиску в них. Суть запропонованої методики й особливості її використання описано на різних прикладах. Подробно описаны вероятностные методики, основанные на обновлении данных по Байесу и расчете надежности конструкций. Методики используются для оптимизации диагностических осмотров и (или) частоты проведения локальних инспекций трубопроводов. Это гарантирует экономичное решение проблем обеспечения работоспособности трубопроводов, продления их долговечности или оптимизации давления в них. Суть предложенной методики и особенности ее применения детально описаны на различных примерах. en Інститут проблем міцності ім. Г.С. Писаренко НАН України Проблемы прочности Научно-технический раздел Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results Продление долговечности трубопроводов и меры по обеспечению их работоспособности, основанные на оптимизированном использовании результатов наземной диагностики и локальной инспекции Article published earlier |
| spellingShingle | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results Francis, A. McCallum, M. Jandu, C. Научно-технический раздел |
| title | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| title_alt | Продление долговечности трубопроводов и меры по обеспечению их работоспособности, основанные на оптимизированном использовании результатов наземной диагностики и локальной инспекции |
| title_full | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| title_fullStr | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| title_full_unstemmed | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| title_short | Pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| title_sort | pipeline life extension and integrity management based on optimized use of above ground survey data and inline inspection results |
| topic | Научно-технический раздел |
| topic_facet | Научно-технический раздел |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/48440 |
| work_keys_str_mv | AT francisa pipelinelifeextensionandintegritymanagementbasedonoptimizeduseofabovegroundsurveydataandinlineinspectionresults AT mccallumm pipelinelifeextensionandintegritymanagementbasedonoptimizeduseofabovegroundsurveydataandinlineinspectionresults AT janduc pipelinelifeextensionandintegritymanagementbasedonoptimizeduseofabovegroundsurveydataandinlineinspectionresults AT francisa prodleniedolgovečnostitruboprovodovimerypoobespečeniûihrabotosposobnostiosnovannyenaoptimizirovannomispolʹzovaniirezulʹtatovnazemnoidiagnostikiilokalʹnoiinspekcii AT mccallumm prodleniedolgovečnostitruboprovodovimerypoobespečeniûihrabotosposobnostiosnovannyenaoptimizirovannomispolʹzovaniirezulʹtatovnazemnoidiagnostikiilokalʹnoiinspekcii AT janduc prodleniedolgovečnostitruboprovodovimerypoobespečeniûihrabotosposobnostiosnovannyenaoptimizirovannomispolʹzovaniirezulʹtatovnazemnoidiagnostikiilokalʹnoiinspekcii |